NAG Ping - Issue 17 (December 2004)

This issue covers:

• Next Generation Data Mining Algorithms Now Available
• Celebrating 50 Years of Fortran
• IRIS Explorer - Release 5.2 is Launched
• Mathematics on the NET
• Tech Tips - Windows Computing from the Microsoft Office Suite
• Product News - New Implementations
• Your privacy - our policy
• Subscribe/Unsubscribe - Manage Your Account
• About The NAG Ping
• Contact us / Copyright information

NAG is pleased to announce that release 2.0 of the NAG Data Mining & Cleaning Components is now available. The NAG Data Mining & Cleaning Components is the first commercially available data mining application development toolkit that uses the results from a three-year EU-funded project, EUREDIT, a collaborative network of more than 300 computer scientists and mathematical experts that work together to solve complex mathematical problems.

Technical improvements at release 2.0 of the NAG Data Mining & Cleaning Components include:

1. Enhanced data cleaning to resolve the problems of missing, invalid or incomplete data (data imputation methods);
2. Advances in outlier identification to determine which datasets are suitable for analysis;
3. Newly memory-efficient multivariate statistical methods that have been the traditional core of data mining techniques; and,
4. A wide range of added functionality for machine learning and pattern recognition.

NAG Data Mining & Cleaning Components, release 2.0, unlike most other commercially available data mining tools, is designed to be incorporated into the user's application rather than requiring the user to learn a new interface to gain access to additional techniques.

To read more about the NAG Data Mining & Cleaning Components click here.
To download a free trial, click here.

It’s been 50 years since John Backus' team produced the very first Fortran compiler. In honor of this anniversary we've asked Malcolm Cohen, Principal Technical Consultant and leader of Fortran compiler development at NAG, to write two articles with a historical perspective.

The opening paragraph of "Fortran: A Few Historical Details" is featured below.

In 1954 computing history was made when an IBM team lead by John Backus produced Fortran, the first high-level programming language. Because its competition was assembly languages, efficiency was a primary concern and so the first compiler was an optimizing compiler. This focus on efficiency has continued throughout Fortran's history.

Read the entire article here. In the next edition of this newsletter we will feature Malcolm's second article entitled "A Personal History of the NAGWare Fortran Compiler".

NAG is pleased to announce the release of version 5.2 of its visualization software, IRIS Explorer. IRIS Explorer 5.2 is a maintenance release which incorporates bug fixes and improved support for a number of its software modules.

IRIS Explorer is a visual programming system for data visualization, animation, manipulation and analysis. It has been designed for software developers, scientists and engineers to create applications for displaying and analyzing complex multi-dimensional datasets interactively.

At its launch, release 5.2 is being made available on the Red Hat 8.0 platform. This improves upon the previous Linux release of IRIS Explorer since the previous release was built under an earlier version of Red Hat and shows some incompatibilities with later versions of the operating system. The next platform for IRIS Explorer 5.2 is scheduled to be 64-bit AMD Opteron. While the Linux build of IRIS Explorer will run on the Opteron in 32-bit mode, a special build is required to take full advantage of the 64-bit architecture. IRIS Explorer 5.2 for Opteron will meet this requirement.

For more information about NAG's IRIS Explorer please visit here.

The EU-funded Mathematics On The NET (MONET) project which finished earlier this year has just received its final review by external experts who awarded it the maximum possible grading of "excellent". The project developed a framework for describing mathematical web services to be used by a person deploying a service to advertise its capabilities, and by a potential client to describe the functionality which he or she requires. A software agent called a broker mediates between the two by discovering which services might potentially meet these needs, and by helping the client to invoke one or more of them.

MONET's service description and discovery mechanism is based on the new Web Ontology Language (OWL) from W3C. This is envisaged as a fundamental technology for building future "semantic webs" -- extensions of the existing worldwide web where information is encoded using mechanisms which allow it to be manipulated by software (such as the MONET broker) in a semantically-meaningful way. The software infrastructure, including the broker and a number of prototype web services, was implemented using emerging web standards such as SOAP and WSDL.

As well as managing MONET, NAG played a leading role in the design and implementation of the OWL ontologies, and implemented and deployed a number of web services based on its world-famous mathematical library. Future work in this area may look at its application to e-Science, and to the rapid assembly of large-scale, distributed applications in HPC. Further information about MONET may be found on the project website at http://monet.nag.co.uk.

The NAG DLL implementations provide an excellent source of algorithmic power to boost the functionality of the Microsoft Office suite. Previous articles have described how easy it is to use the libraries in this way. The DLL libraries are provided with examples of use within the Excel package and also the necessary interfacing software to VB and VBA.

Some users have asked why when they solve a problem within Excel using a NAG routine, slightly different answers are obtained when they repeat the same calculation immediately afterwards. The Intel chip can support 80-bit and 64-bit operation. The NAG libraries use the 64-bit operating mode because this is the default operation with the compilers used to form the DLLs. The first time a problem is solved this will be the mode of operation. The mode can be changed by setting the Floating Point Control Word, and Excel does this to set the mode of working to 80-bit. In consequence the second time the problem is solved, the NAG libraries will start the calculation in 80-bit mode.

For well-conditioned problems this is not serious and the NAG library routines will switch back to 64-bit working whenever it is essential for the correct working of an algorithm, but it can be disconcerting to see this behavior. A solution is to set the Floating Point Control Word before each call to a NAG routine. At the next release of the Library, NAG will expose this functionality from the DLLs when an external compiler function is required.

For more information on NAG's DLLs click here.

NAG C Library, Mark 7

• Intel-32 Linux using the Intel C++ v 8.0 compiler
• Intel-32 Linux using the gcc v 3.2.2 compiler
• Intel-64 Linux64 using the Intel C++ v 7.0 compiler
• Intel-64 Linux64 using the Intel C++ v 8.0 compiler
• Intel-64 Linux64 using the gcc v 3.2.3 compiler

NAG Fortran Library, Mark 20

• Intel-64 Linux64 using the Intel v 8.0 compiler

IRIS Explorer, Release 5.2

• Intel-32 Linux

NAG DMC (Data Mining and Cleaning Components), Release 2

• HP Alpha Tru64 using the Tru64 Compiler
• Intel-32 Windows DLL using the MSV C/C++ v 6.0 compiler
• Intel-32 Linux using the gcc v 3.2.2 compiler
• Sun SPARC Solaris using the Workshop 5 C compiler
• Apple Mac OS X

NAGWare f95 compiler, Release 5.0

• SGI MIPS IRIX 6
• Intel-64 Linux 64
• HP PA-RISC 1.1 HP-UX11 using the HP ANSI C compiler
• Sun SPARC Solaris using the gcc 2.95.2 compiler
• AMD-64 Linux64 using the gcc 3.2.1/3.3.1 compiler
• HP Alpha Tru64
• IBM Power AIX using the AIX SL C 3.1.4.0 compiler

NAG Foundation Toolbox, Release 1.1

• Sun SPARC Solaris for MATLAB 6.1, 6.5 or 7

At NAG, we have a simple policy on the use of information you provide to us; it is to treat you how we would like to be treated. We will not sell or give your personal information to anyone and will only use it for the purposes we state and with the permission you grant us. NAG also is one of the first U.S. companies to voluntarily certify compliance with the safe harbor principles for privacy protection developed by the U.S. Department of Commerce in conjunction with the European Union.
To learn more, click here.

You have received NAG Ping because you have:

1. Registered on the NAG Web site
2. Hold a current license to NAG Software
3. Given NAG permission to send information to you from a prior research survey
4. Requested NAG add you to their mailing list

The NAG Ping is an occasional newsletter of technical tips, product information and discussions of technical computing topics. Its principal audience is in North America though it has readers throughout the world. It is the sister publication of the NAGNews, produced by our colleagues at NAG Ltd primarily for those in Europe and elsewhere outside North America.

The NAG Ping is named in memory of Mike Muuss, formerly of the Army Research Laboratory. Among the creative and useful works of his life was the "ping" program to test network connection. It was written by Mike in 1983 to help diagnose problems on his network. The name is tied to Mike's earlier work in sonar modeling and echolocation. To learn more, click here.

Contact us at at ping@nag.com or http://nag.com/contact_us.asp#northamerica.

Copyright 2001-2004 Numerical Algorithms Group, Inc.